THERMOMETER FOR THE DETERMINATION OF RELATIVE HEIGHTS. 
127 
The discrepancies in many of the above results as compared with the column of 
previously determined heights are considerable, but it is to be observed that in most 
cases where the height of the station may be considered as particularly well-esta- 
blished, the observed height accords with it in a very remarkable manner. This is 
particularly the case at Geneva (where the height is of course deduced only from that 
of the Great St. Bernard), at Bonneville, at Chamounix (taking a mean of the six 
observations, which gives 3442), at the Croix du Bonhomme*, at St. Remy, at the 
Great St. Bernard and at Brussone : the chief exceptions being at the Mole-f-, the 
Montanvert, the Col di Val Dobbia, Orta, Mogadino, Domo d’Ossola, the Simplon 
Hospice and Brieg, the great differences at which places are to me quite unaccount- 
able, unless on the supposition, in the case of the five last-named, that their distances 
from the fixed stations were too great to allow of trustworthy results being obtained 
from single observations. In confirmation of this view, it will be seen presently that 
these observations give far more satisfactory results when employed without any 
reference to the simultaneous observations at Geneva and the Great St. Bernard ; and 
also in the above Table we may notice that, after crossing the Bonhomme, in those 
cases where the heights determined from the two fixed stations differ from one 
another to any considerable amount, that obtained by reference to the Great St. 
Bernard generally approximates more nearly to the height as previously determined 
than the other, obtained by a comparison with the more distant station. 
The numbers in the tenth column which are the sums of the two values, the height 
above Geneva and the depression below the Great St. Bernard, as determined by ob- 
servation, should, supposing both theory and observations to be perfect, give the con- 
stant value 6836, which is the difference of height of those two stations ; but this 
difference will vary when determined by means of each pair of barometric observa- 
tions, the standard value 6836 being the mean of a very considerable number. Now 
it is very evident that were it not for that term in the formula which involves the 
correction for the expansion of the air, the sum of the above values would give the 
difference of altitude, however erroneous the intermediate observation might be : for 
calling b', b" and b 1 " the three boiling-points, we should have 
548(Z/ — &") — 24D + 548(&" — 6"')-j-24D = 548(6'— £/"). 
The observation, therefore, of the external temperature at the intermediate station 
becomes the means of rendering the accordance of the sums of the component alti- 
tudes with the whole height between Geneva and the Great St. Bernard, as deter- 
mined in each case, a test of the correctness of the observation of the boiling tempe- 
rature. We may in fact, by equating the formula for the whole height with the sum, 
* All who have traversed the Pass of the Bonhomme are aware that in crossing to the south, upon Chapiu, 
the tract rises considerably after passing what ought to be and, I believe, is considered the “ Col,” viz. the point 
from which a view of the mountains of the Tarentaise is first obtained ; this point is marked by a cross, larger 
than usual on most passes, and than any of the numerous others on this pass. This cross I believe to be the 
“ Croix du Bonhomme” of De Saussure, and it was at its base the observation above referred to was made. 
f See note, p. 131 . 
MDCCCXLVI. S 
